After Gen upgrade from 6.0 to 8.5 an application that runs under the Traditional Chinese/Big5 Windows locale has some characters corrupted after flow from the Gen client to the Gen z/OS server and return. These are not User Defined Characters (UDC).
Post Gen 6.0 GA there was a change to the content of the Gen codepage translation tables for Windows Big5 (codepage 950) <-> IBM z/OS Traditional Chinese (codepage 937) because of a switch from using the IBM to Microsoft codepage on the Windows OS.
The problems can be circumvented by reverting to the 6.0 GA translation tables for 950 <-> 937 on the z/OS server and rebuilding TIRCRUNC.
The 2 table names are as follows:
ADKJADLG (937 -> 950)
ADLGADKJ (950 -> 937)
A copy of both 6.0 GA tables is attached here: 60_GA_950_937_tables.zip
This problem will be investigated further by Engineering to be fixed in a future release of Gen